Quantitative trait locus mapping of yield and yield components in canola (Brassica napus L.) under irrigated and rainfed treatments

Abstract

Rapeseed (Brassica napus L.) is an oilseed crop that has a variety of uses, including applications in industry as well as for food, feed, and fuel. Improved B. napus cultivars with decreased levels of two disease-causing compounds are known commonly as canola or double-low cultivars, and are approved for human and animal consumption in the U.S., Canada, and Europe. Canola is currently grown in the northern U.S. and in several areas of Canada and Europe, but has potential to be grown in other areas of the U.S., including Colorado, either for biodiesel or to be sold in a canola commodity food oil market at the discretion of the seed producer. Additionally, the cake meal left after oil extraction has a high protein concentration and can be added as a supplement to animal feed. In recent years, water availability for crop production in the western U.S. has declined due to competition with non-agricultural water uses, and the increasing demands for water will likely increase with global climate change. Therefore, in order to be sustainable, crops grown in Colorado must be high-yielding with limited or no irrigation inputs, and consequently canola cultivars adapted to the semi-arid climate of Colorado and the U.S. High Plains will need to be drought tolerant. To provide information relevant to improving adaptation of canola to Colorado conditions, a study was conducted with the following objectives: 1) to evaluate yield, yield components, and days to flowering (DTF) in two doubled haploid (DH) canola mapping populations under rainfed and irrigated conditions; 2) to determine relationships among yield and yield components by analyzing trait correlations and to study trait inheritance patterns; 3) to determine areas of the B. napus genome that are implicated in yield and yield component traits under both rainfed and irrigated conditions by quantitative trait locus (QTL) analysis; and 4) to study the sensitivity of yield and yield component traits to drought stress by performing analysis of variance and by performing a QTL analysis on the difference in trait values from the rainfed and irrigated treatments. Two DH canola mapping populations were grown in side-by-side irrigated and rainfed treatments near Fort Collins, Colorado: population SE1 in 2010 (n=183) and population DHYB (n=150) in 2011. DTF, seed yield, and yield-related traits were measured in order to understand relationships among these traits under different water regimes, to study trait heritabilities, and to better understand genotype, treatment, and treatment by genotype interaction effects. QTL mapping was conducted separately for each treatment in each population using R-QTL software to detect additive and epistatic effects. Yield components that were studied included siliques per main inflorescence (SMI), seeds per silique (SS), and thousand seed weight (TSW). Seed coat color was also classified for the DHYB population. Analysis of variance revealed an influence of genotype (P<0.0001) on all traits in both populations, treatment effects on seed yield, SMI, and SS (P<0.05) in the SE1 population, and treatment effects on seed yield, SMI, TSW, and DTF in the DHYB population. Genotype by treatment interactions were significant (P<0.01) for all traits in the SE1 population and for seed yield and TSW (P<0.05) in the DHYB population. In the 2010 study, three DTF QTL were detected that colocated with most of the other QTL, demonstrating the strong influence of flowering time on seed yield and yield components in this population. These QTL explained 73 and 65 percent of phenotypic variance for DTF in the wet and dry treatments, respectively, in a multiple QTL model. Several novel QTL were detected in the 2011 study, including a locus on LG 17 that explained 7.54% of trait variation for seed yield in the dry treatment and a locus on LG 16 that explained 11.41% of seed yield in the wet treatment. A novel QTL for SS was detected on LG 7 in both treatments. Two QTL reported for DTF in the wet treatment on LGs 14 and 18 are novel, as are two QTL for SMI in the dry treatment. Of the yield components studied, SS consistently had the lowest amount of genotype by treatment interaction and direct treatment effects in both populations, as well as high heritability estimates in both populations. SS also correlated positively and significantly (P<0.05) with seed yield in both years of the study. Additionally, the QTL detected for SS in the 2011 study were the same for the wet and dry treatments, indicating QTL stability. It is our conclusion that SS is a good candidate for direct selection in a breeding program, and that he QTL reported for SS could also be useful for marker-assisted selection for improved yields in Colorado.